跳至内容

AWS Textract : c. is always returned as C.

0

AWS textract is superb and gives good results for documents which contains images. However what we observed is that everytime there is c. like in attached image Sample Image

It always returns result as C. (In capital). Rest of dots like a. b. d. e. etc never had issues. Observed this with multiple documents.

result rawText.txt set of above image which was part of PDF

Right of Citizen, Political Party, Organization or Association To canvass for votes in an Election 82. a. Right of Citizen, Association or Organization To Contribute Funds to Political Party or Independent Candidate b. Prohibition on Possession of Funds Outside the Republic or Contribution from Outside the Republic to any Political Party or Independent Candidate C. Power of Elections Commission to Examine and Audit the Books of Political Parties or Independent Candidates 83. a. time for Conducting General Elections b. Elections of Public Officers to be Determined by Absolute Majority C. Power of Elections Commission to Declare Results of Elections and Right to Appeal by any Aggrieved Party or Candidate d. Liabilities by Political Parties and Independent Candidates

已提问 2 年前449 查看次数

1 回答
1

Hi, thank you for using AWS Textract. Unfortunately, this is a model quality issue as this is a hard case in this font to know if that's a c or a C in using the word-level information. We will try to improve this in our upcoming model releases.

AWS

已回答 2 年前

  • Thank you for the reply. Look forward for the fix in future releases.

您未登录。 登录 发布回答。

一个好的回答可以清楚地解答问题和提供建设性反馈,并能促进提问者的职业发展。